226389 Analyzing Longitudinal Data Via Multivariate Individual Change Models: Applications in Underserved Populations

Tuesday, November 9, 2010 : 8:30 AM - 8:50 AM

Joseph Rausch, PhD , Division of Behavioral Medicine and Clinical Psychology, Cincinnati Children's Hospital and Medical Center, Cincinnati, OH
     
Longitudinal data provide a unique opportunity for documenting links among important outcomes. Although investigators can choose from a variety of analytic methods, multivariate individual change models (MICMs) provide an especially worthwhile approach for analyzing and presenting associations among multiple variables and their rates of change over time. In particular, these models provide a number of advantages over traditional methods (e.g., repeated measures ANOVA and MANOVA) and provide a complex, yet comprehensible, understanding of relevant public health outcomes, especially in underserved populations.

In this talk, I will present MICMs from a multilevel modeling perspective to emphasize the nested structure inherent in longitudinal data. I will describe the general modeling framework, including how one can evaluate questions concerning associations between rates of change in multiple variables and bidirectional effects (also known as reciprocal or cross-lagged effects). I will address specific research questions concerning individual and group change via the parameters from the statistical model, with particular emphasis on underserved populations in the health care community. I will briefly describe the application of MICMs with statistical software packages and provide an illustrative data example to make the core concepts concrete for applied researchers and policy makers. This example will demonstrate (a) the utility of the method for analyzing longitudinal data on multiple variables in an underserved population and, more generally, (b) the importance of seriously considering MICMs for the analysis of longitudinal data in public health research.

Learning Areas:
Biostatistics, economics

Learning Objectives:
At the session’s end participants will be able to: 1. Compare multivariate individual change models to more traditional approaches for the analysis of longitudinal data (e.g., repeated measures ANOVA and MANOVA). 2. Accurately explain the meaning of the parameters in the multivariate individual change model and their relevance to public health research using a specific example. 3. Identify two specific settings in public health research where multivariate individual change models could be employed to better understand difficulties in underserved populations.

Keywords: Child/Adolescent, Methodology

Presenting author's disclosure statement:

Qualified on the content I am responsible for because: I have published in well-known methodological journals including Psychological Methods, Applied Psychological Measurement, and the Annual Review of Psychology and have delivered more than 20 research presentations on statistical methods and their application, focusing primarily on longitudinal methods for the analysis of change. For this presentation, I worked individually to obtain the pertinent background information, summarize the general framework for the presented methodology and its relevance to public health research, and analyze the data in the illustrative example.
Any relevant financial relationships? No

I agree to comply with the American Public Health Association Conflict of Interest and Commercial Support Guidelines, and to disclose to the participants any off-label or experimental uses of a commercial product or service discussed in my presentation.

Back to: 4067.0: Statistical Modelling for Social Justice